Scale-Dependent Interactions in Insect Food Web Modules

Species interactions likely produce scale-dependent patterns due to the dispersal of different species. Understanding the scale-dependency of species interaction is critical to the development of ecology theories, and is one necessary step toward better extrapolating experimental results to landscape-scale applications. I studied the scale-dependency of species interactions on food web modules based on a salt marsh plant (marsh elder, Iva frutescens). (1) I explored how predator-prey interactions between ladybeetles and aphids vary with spatial scale, using field population data, a field experiment and a spatially-explicit model. I found that both top-down and bottom-up effects were more striking at smaller scales. Using post hoc multi-scale analysis, I could confirm that the scale-dependency patterns found here were not caused by experimental artifacts. (2) I used enclosure experiment, field time-series data and a structural equation model based on two-month average population data to study side-to-side interactions between three herbivorous insects that all feed on Iva frutescens. From the field time series data, I found that many side-to-side interactions between herbivorous insects were only detected when observations were made with certain time scales. Theses delayed interactions cannot be found with experiment that focused on relatively short-term effects. Finally, (3) I studied the arthropod food web based on Iva frutescence at a much larger spatial scale and examined latitudinal autocorrelation patterns in the arthropod communities by analyzing field survey data. By comparing beta-diversity patterns of morpho-species with beta-diversity patterns of guilds, I concluded that geographic changes in community composition were best explained by changes in food web structure, rather than environmental filtering across the latitudinal gradient. In sum, my results suggest that observed patterns of species interactions vary a lot depending on the spatial and temporal scales at which observations are made, and the empirical approaches through which interactions were documented. My work illustrates that using multi-scale analysis and multiple methods to examining the same system allows discerning the effect of scale per se. These approaches would lead to deeper insights into the logical links between processes and patterns of species interactions.Biology and Biochemistry, Department o